Stable Diffusion 大模型资源下载集合

0xSoul · 发表于 2023-4-17 10:10:33

Stable Diffusion

Stable Diffusion v2.0

Stable Diffusion 2.0 源码 https://github.com/Stability-AI/stablediffusion

768x768 模型 https://huggingface.co/stabilityai/stable-diffusion-2

512x512 基础模型 https://huggingface.co/stabilityai/stable-diffusion-2-base

深度模型 https://huggingface.co/stabilityai/stable-diffusion-2-depth

4倍放大器 https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler

Stable Diffusion v1.5 [81761151] [a9263745]

“Stable-Diffusion-v1-5”checkpoint是使用“Stable-Diffusion-v1-2”检查点的权重进行初始化的，随后在“laion-aesthetics v2 5+”数据集上进行了512x512分辨率的595k步微调，并且减少了10％的文本条件，以改善无分类器的引导采样。仓库地址为 [https://huggingface.co/runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5)。

!!! 注释 v1-5-pruned-emaonly.ckpt [81761151] | SHA256 cc6cb27103417325ff94f52b7a5d2dde45a7515b25c255d8e396c90014281516 v1-5-pruned.ckpt [a9263745] | SHA256 e1441589a6f3c5a53f5f54d0975a18a7feb7cdf0b0dee276dfc3331ae376a053

磁力链下载 v1-5-pruned-emaonly.ckpt

magnet:?xt=urn:btih:2daef5b5f63a16a9af9169a529b1a773fc452637&dn=v1-5-pruned-emaonly.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2ftracker.openbittorrent.com%3a6969%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker2.dler.org%3a80%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zemoj.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.lelux.fi%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.army%3a6969%2fannounce

Web下载 v1-5-pruned-emaonly.ckpt

[https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt](https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt)

磁力链下载 v1-5-pruned.ckpt !!! 警告此版本的模型仅适用于原生训练。如果您不知道这是什么，您就不需要使用此版本。如果您只想生成图像或训练Dreambooth / 文本反演 / 超网络，请使用v1-5-pruned-emaonly.ckpt。

magnet:?xt=urn:btih:38c2b7a8abf55604ae8a5f060ffda2e028075af3&dn=v1-5-pruned.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2ftracker.openbittorrent.com%3a6969%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker2.dler.org%3a80%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zemoj.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.lelux.fi%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.army%3a6969%2fannounce

Web下载 v1-5-pruned.ckpt Requires login. !!! 警告此版本的模型仅适用于原生训练。如果您不知道这是什么，您不需要使用此版本。如果您只想生成图像或训练Dreambooth / 文本反演 / 超网络，请改为使用v1-5-pruned-emaonly.ckpt。

[https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned.ckpt](https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned.ckpt)

Stable Diffusion VAE ft-mse-original

从ft-EMA恢复并使用EMA权重，使用重新加权的损失进行另外280k步训练，更加重视MSE重建（产生了稍微“平滑”一些的输出）。为了保持与现有模型的兼容性，只有解码器部分进行了微调；这些检查点可以用作现有自编码器的即插即用替代品。仓库及信息（包括比较）：[https://huggingface.co/stabilityai/sd-vae-ft-mse-original](https://huggingface.co/stabilityai/sd-vae-ft-mse-original)

!!! 警告 VAE与列表中的其他模型不同。它不是一个可以单独用于生成图像的完整模型，而是一个与其他模型配合使用的组件。如果您不知道这是什么或者为什么需要它，请不要理会它。

Web下载

[https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt](https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt)

Stable Diffusion VAE ft-ema-original

从原始检查点中恢复，训练了313198步，并使用EMA权重。为了保持与现有模型的兼容性，只有解码器部分进行了微调；这些检查点可以用作现有自编码器的即插即用替代品。仓库及信息（包括比较）：[https://huggingface.co/stabilityai/sd-vae-ft-ema-original](https://huggingface.co/stabilityai/sd-vae-ft-ema-original)

!!! 警告 VAE与列表中的其他模型不同。它不是一个可以单独用于生成图像的完整模型，而是一个与其他模型配合使用的组件。如果您不知道这是什么或者为什么需要它，请不要理会它。

Web

[https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt](https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt)

Stable Diffusion Inpainting v1.5 [3e16efc8]

仓库链接 [https://huggingface.co/runwayml/stable-diffusion-inpainting](https://huggingface.co/runwayml/stable-diffusion-inpainting) 旨在与 [https://github.com/runwayml/stable-diffusion](https://github.com/runwayml/stable-diffusion) 一起使用。

Stable-Diffusion-Inpainting 是用 Stable-Diffusion-v-1-2 的权重初始化的。首先进行了 595k 步的常规训练，然后进行了 440k 步的修复训练，在“laion-aesthetics v2 5+”上的分辨率为512x512，并减少了10%的文本调节，以改进无分类器指导采样。对于修复，UNet具有5个额外的输入通道（4个用于编码的遮罩图像，1个用于遮罩本身），在恢复非修复检查点后进行零初始化权重。在训练期间，我们生成合成遮罩，并在25%的情况下对所有内容进行遮罩。

!!! 注释 SHA256 c6bbc15e3224e6973459ba78de4998b80b50112b0ae5b5c67113d56b4e366b19

Torrent

magnet:?xt=urn:btih:b523a9e71ae02e27b28007eca190f41999c2add1&dn=sd-v1-5-inpainting.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2ftracker.openbittorrent.com%3a6969%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker2.dler.org%3a80%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zemoj.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.swateam.org.uk%3a2710%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.lelux.fi%3a6969%2fannounce

Web Requires login

[https://huggingface.co/runwayml/stable-diffusion-inpainting/resolve/main/sd-v1-5-inpainting.ckpt](https://huggingface.co/runwayml/stable-diffusion-inpainting/resolve/main/sd-v1-5-inpainting.ckpt)

Stable Diffusion v1.4 [4af45990] [7460a6fa] [06c50424]

!!! 注释 sd-v1-4-pruned.ckpt [4af45990] sd-v1-4.ckpt [7460a6fa] | SHA256 fe4efff1e174c627256e44ec2991ba279b3816e364b49f9be2abc0b3ff3f8556 sd-v1-4-full-ema.ckpt [06c50424] | SHA256 14749efc0ae8ef0329391ad4436feb781b402f4fece4883c7ad8d10556d8a36a

磁力链

magnet:?xt=urn:btih:3A4A612D75ED088EA542ACAC52F9F45987488D1C&tr=udp://tracker.opentrackr.org:1337

HuggingFace Requires login.

[https://huggingface.co/CompVis/stable-diffusion-v-1-4-original](https://huggingface.co/CompVis/stable-diffusion-v-1-4-original)

GDrive

[https://drive.google.com/file/d/1wHFgl0ivCmIZv88hVZXkb8oy9qCuaBGA/view](https://drive.google.com/file/d/1wHFgl0ivCmIZv88hVZXkb8oy9qCuaBGA/view)

Waifu Diffusion

Waifu Diffusion VAE kl-f8-anime2

对 SD 1.4 VAE 模型在 25 万张动漫风格图像上进行了微调，据说可以改善眼睛和手指的重建效果。信息 [https://twitter.com/haruu1367/status/1579286947519864833](https://twitter.com/haruu1367/status/1579286947519864833) 存储库https://huggingface.co/hakurei/waifu-diffusion-v1-4`

！！！警告 VAE 不同于列表中的其他模型。它不是可以单独用于生成图像的完整模型，而是与其他模型配合使用的组件。如果您不知道这是什么或为什么需要它，请不要使用它。

Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime2.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime2.ckpt)

Waifu Diffusion VAE kl-f8-anime

一种在250k个动漫风格图像上对SD 1.4 VAE进行微调的方法，据称可以改善眼睛和手指的重建。信息在 [https://twitter.com/haruu1367/status/1579286947519864833](https://twitter.com/haruu1367/status/1579286947519864833`)，代码在 [https://huggingface.co/hakurei/waifu-diffusion-v1-4](https://huggingface.co/hakurei/waifu-diffusion-v1-4`)。

!!! 警告 VAE不同于列表中的其他模型。它不能单独用于生成图像，而是要与其他模型一起使用的组件。如果您不知道这是什么或者为什么需要它，请不要费心。

Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime.ckpt)

Waifu Diffusion v1.3 Full [84692140] [4470c325] [e1de58a9] [3e1a125f]

该模型是在600,000个高分辨率Danbooru图像上进行了10个Epoch的训练的1.3 Waifu Diffusion模型。发布说明可以在此链接中找到：[https://gist.github.com/harubaru/f727cedacae336d1f7877c4bbe2196e1](https://gist.github.com/harubaru/f727cedacae336d1f7877c4bbe2196e1)

!!! note wd-v1-3-float16.ckpt [84692140] | SHA256 4afab9126057859b34d13d6207d90221d0b017b7580469ea70cee37757a29edd wd-v1-3-float32.ckpt [4470c325] | SHA256 9dade826203c7ee369881b5dc20d34298fa644c1f137568e09fbc4b9a0d3e817 wd-v1-3-full.ckpt [e1de58a9] | SHA256 23ba8d0411c211d3d14903d46613bc088924e1453ed1c6428ce86bde54a37d27 wd-v1-3-full-opt.ckpt [3e1a125f] | SHA256 10912b9a6d773ea7c299c0563d10538ada04ade81837362b6c0c67be4df937c1

wd-v1-3-float16.ckpt [84692140] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-float16.ckpt)

wd-v1-3-float32.ckpt [4470c325] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-float32.ckpt)

wd-v1-3-full.ckpt [e1de58a9] Web 这里提到的完整EMA比较大是因为它包含了仅用于训练的附加数据（training only）。如果您不知道这是什么，请使用float16或float32版本即可。

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-full.ckpt)

wd-v1-3-full-opt.ckpt [3e1a125f] Web 完整的opt文件大小较大，因为它包含仅用于训练的其他数据（training only）。如果您不知道这是什么，请使用float16或float32版本

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-full-opt.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/wd-v1-3-full-opt.ckpt)

Waifu Diffusion v1.3 beta epoch09 [a0d90b41] [134eaffb] [e90a9d02]

==未完成的Beta版本== 使用600,000张高分辨率Danbooru图像在9个Epochs上训练的1.3 Waifu Diffusion模型。

model-epoch08-float16.ckpt [a0d90b41] | SHA256 b30bab13a2ea946ba85bc7b3a30c3fa20d013ca306193aa54b4771d147192dc1 model-epoch08-float32.ckpt [134eaffb] | SHA256 a16d2f57b229156a61ccab6b1cbc120332f3a2a320c75952241e6439f3fe79ec model-epoch08-full.ckpt [e90a9d02] | SHA256 7fb9c64bb460724dfecbf693fcd8e0e556c4664553a352c1256997813d585329

model-epoch09-float16.ckpt [a0d90b41] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-float16.ckpt)

model-epoch09-float32.ckpt [134eaffb] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-float32.ckpt)

model-epoch09-full.ckpt [e90a9d02] Web The full EMA is larger in file size because it contains additional data used for training only. If you don't know what this is, use the float16 or float32 version.

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch09-full.ckpt)

Waifu Diffusion v1.3 beta epoch08 [55537280] [7dd9e8e1] [d12b4159]

==未完成测试版== 基于60万张高分辨率Danbooru图像训练的1.3 Waifu Diffusion模型，训练了8个Epoch。

!!! note model-epoch08-float16.ckpt [55537280] | SHA256 3a3a36aac85502a06878ec6760c0fb95f71546ad734fbb2b6cfab9a50db7aa8d model-epoch08-float32.ckpt [7dd9e8e1] | SHA256 e64cbc29edbef82142768c3a398345f602ba8b337ebd22cd83071162fe789a18 model-epoch08-full.ckpt [d12b4159] | SHA256 c2a8b09f77ce572c012c810495d8ae9f989d04ca6b3226c93be5dce2208bacdd

model-epoch08-float16.ckpt [55537280] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-float16.ckpt)

model-epoch08-float32.ckpt [7dd9e8e1] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-float32.ckpt)

model-epoch08-full.ckpt [d12b4159] Web “full EMA” 的文件大小更大，因为它包含了仅用于训练的额外数据。如果您不知道这是什么，请使用 float16 或 float32 版本。

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch08-full.ckpt)

Waifu Diffusion v1.3 beta epoch07 [3ad68e9a] [e4150f29] [39312118]

==未完成测试版== 在600,000张高分辨率的Danbooru图像上进行了7个Epoch的训练的1.3 Waifu Diffusion模型。

该模型还没有完成，离完成还很远。不要将结果视为最终结果。

!!! note model-epoch07-float16.ckpt [3ad68e9a] | SHA256 0aa0a2d1620b65fbcfe800ac5d5b5fc681c9e3bc7373ad5519837f94828eb83f model-epoch07-float32.ckpt [e4150f29] | SHA256 99c1061162c126c59d17cfddaf8b475935a9f0a504f7b305b1d733e7618be82b model-epoch07-full.ckpt [39312118] | SHA256 74ae340acb9ac251b8b9ee57f795e18bc4d39c016272fafd440e637fe391a790

model-epoch07-float16.ckpt [3ad68e9a] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-float16.ckpt)

model-epoch07-float32.ckpt [e4150f29] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-float32.ckpt)

model-epoch07-full.ckpt [39312118] Web The full EMA is larger in file size because it contains additional data used for training only. If you don't know what this is, use the float16 or float32 version.

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch07-full.ckpt)

Waifu Diffusion v1.3 beta epoch06 [281c28bb] [3a672a8a] [d970eaf1]

==未完成的测试版== 在 600,000 张高分辨率 Danbooru 图像上训练了 6 个 Epochs 的 1.3 Waifu Diffusion 模型。

该模型还没有完成，离完成还很远。不要将结果视为最终结果。

!!! note model-epoch06-float16.ckpt [281c28bb] | SHA256 c0397df53cd4cf1cfadb95a121a67eaad903e4203d708abfe3409a832a517191 model-epoch06-float32.ckpt [3a672a8a] | SHA256 f34c89c80e91d77c06df178d292611ac7757728dd7a3aaff8c4c9117e6e92e35 model-epoch06-full.ckpt [d970eaf1] | SHA256 cc5569691c843f64cc8e819b78c0081f6e8024432f6de898f11ee816ac3ff5ca

model-epoch06-float16.ckpt [281c28bb] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-float16.ckpt)

model-epoch06-float32.ckpt [3a672a8a] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-float32.ckpt)

model-epoch06-full.ckpt [d970eaf1] Web The full EMA is larger in file size because it contains additional data used for training only. If you don't know what this is, use the float16 or float32 version.

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch06-full.ckpt)

Waifu Diffusion v1.3 beta epoch05 [25f7a927] [3563d59f] [9453d3e5]

==未完成测试版== 1.3 Waifu Diffusion 模型使用 60 万高分辨率 Danbooru 图像进行了 5 个 Epochs 的训练。

该模型还没有完成，离完成还很远。不要将结果视为最终结果。

!!! note model-epoch05-float16.ckpt [25f7a927] | SHA256 26cf2a2e30095926bb9fd9de0c83f47adc0b442dbfdc3d667d43778e8b70bece model-epoch05-float32.ckpt [3563d59f] | SHA256 694617b2145abcd248d78ea2c556af86787ef9f79e9d9a0f7473d7b408dd3dfa model-epoch05-full.ckpt [9453d3e5] | SHA256 e0fceaac42328912ed60957c5fed129a418bdd5f3ebe82d5f09e7901a296a6ce

model-epoch05-float16.ckpt [25f7a927] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-float16.ckpt)

model-epoch05-float32.ckpt [3563d59f] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-float32.ckpt)

model-epoch05-full.ckpt [9453d3e5] Web The full EMA is larger in file size because it contains additional data used for training only. If you don't know what this is, use the float16 or float32 version.

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch05-full.ckpt)

Waifu Diffusion v1.3 beta epoch04 [958859c3] [84d80299] [994d2e0f]

==未完成的Beta版本== ，使用了600,000张高分辨率的Danbooru图片训练的1.3 Waifu Diffusion模型，训练了4个Epochs。

该模型还没有完成，离完成还很远。不要将结果视为最终结果。

!!! note model-epoch04-float16.ckpt [958859c3] | SHA256 3fd4bbb7ca76c806d88061f0ce994b763693369bd6ea37863c93c1c6b9d25232 model-epoch04-float32.ckpt [84d80299] | SHA256 5237503095524512c7612f81ef786036aa001fd76693064fb00478115659a335 model-epoch04-full.ckpt [994d2e0f] | SHA256 eee5a4391d6074f08bc6c3de3457c245252bde101e3e651c27b5cf85d4926b08

model-epoch04-float16.ckpt [958859c3] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-float16.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-float16.ckpt)

model-epoch04-float32.ckpt [84d80299] Web

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-float32.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-float32.ckpt)

model-epoch04-full.ckpt [994d2e0f] Web The full EMA is larger in file size because it contains additional data used for training only. If you don't know what this is, use the float16 or float32 version.

[https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-full.ckpt](https://huggingface.co/hakurei/waifu-diffusion-v1-3/resolve/main/model-epoch04-full.ckpt)

Anything-V3.0 [38c1ebe3] [1a7df6b8] [6569e224]

动漫模型信息 [https://www.bilibili.com/read/cv19603218](https://www.bilibili.com/read/cv19603218)

Torrent fp16 [38c1ebe3]

magnet:?xt=urn:btih:45cd353ac4fa87098db5e3a6a349539710a3a1f5&dn=Anything-V3.0-fp16.zip&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zerobytes.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.swateam.org.uk%3a2710%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.encrypted-data.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.army%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce

Torrent fp32 [1a7df6b8]

magnet:?xt=urn:btih:d9db662ab5ace77004b3348c23c9381380c27156&dn=Anything-V3.0-fp32.zip&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zerobytes.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.swateam.org.uk%3a2710%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.encrypted-data.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.army%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce

Torrent Full EMA [6569e224]

magnet:?xt=urn:btih:80460036625fb61dce4bc6e7dab744744309a2a0&dn=Anything-V3.0-fullema.zip&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2fvibe.sleepyinternetfun.xyz%3a1738%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.zerobytes.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.theoks.net%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.swateam.org.uk%3a2710%2fannounce&tr=udp%3a%2f%2ftracker.publictracker.xyz%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.monitorit4.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.encrypted-data.xyz%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.army%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce

Cafe Unofficial Instagram TEST 模型 [50b987ae]

该模型是在大约14万张Instagram图片上进行训练的，这些图片主要来自日本的账户（包括cosplay、模特和个人账户）。虽然这个模型本身可以创造出一些逼真的Instagram风格的图片，但为了发挥其全部潜力，建议将其与另一个模型（例如berry或其他模型）合并使用。

Torrent

magnet:?xt=urn:btih:eb085b3e22310a338e6ea00172cb887c10c54cbc&dn=cafe-instagram-unofficial-test-epoch-9-140k-images-fp32.ckpt&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A80&tr=udp%3A%2F%2Fopentor.org%3A2710&tr=udp%3A%2F%2Ftracker.ccc.de%3A80&tr=udp%3A%2F%2Ftracker.blackunicorn.xyz%3A6969&tr=udp%3A%2F%2Ftracker.coppersurfer.tk%3A6969&tr=udp%3A%2F%2Ftracker.leechers-paradise.org%3A6969

Trinart

trinart_characters_19.2m_stable_diffusion_v1 [d64dab7f] + VAE

这是一个基于stable diffusionv1的模型，它使用约1920万个动漫/漫画风格图像（包括预滚动增强图像）进行训练，并使用大约5万个图像进行最终微调。该模型在给定SDv1模型规范下，在艺术风格的多样性和解剖质量之间寻求一个最佳平衡点。这是与AI小说家/TrinArt服务相同的版本1模型，该服务从9月初到10月14日推出。数据集已经过滤，排除了NSFW或不安全的内容。

!!! note trinart_characters_it4_v1.ckpt [d64dab7f] SHA256 d58058f2c71612aa7492d63ad8e6c32b3159494dc51a29ffe71937cdc49b6681 autoencoder_kl-f8-trinart_characters.ckpt SHA256 d2dd1c82220e31a72bd9958dda249ed7f94faf875d5123ae3aab7a1950a82a8f

Web

[https://huggingface.co/naclbit/trinart_characters_19.2m_stable_diffusion_v1/resolve/main/trinart_characters_it4_v1.ckpt](https://huggingface.co/naclbit/trinart_characters_19.2m_stable_diffusion_v1/resolve/main/trinart_characters_it4_v1.ckpt)

VAE Web

[https://huggingface.co/naclbit/trinart_characters_19.2m_stable_diffusion_v1/resolve/main/autoencoder_kl-f8-trinart_characters.ckpt](https://huggingface.co/naclbit/trinart_characters_19.2m_stable_diffusion_v1/resolve/main/autoencoder_kl-f8-trinart_characters.ckpt)

trinart2_step115000.ckpt [f1c7e952]

该 V2 checkpoint使用了 Dropout 技术，在训练数据上增加了 10,000 张图片，并采用了一种新的标记策略，训练时间更长，以提高结果的同时保留原始美学。 !!! note SHA256 776af18775dfccf29725a994df855e7d8f7b8ea525013e3a466f210ec15c8fd4

Web

[https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step115000.ckpt](https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step115000.ckpt)

trinart2_step95000.ckpt [cf0bd941]

V2 checkpoint使用了dropout，增加了1万张图像，并采用了新的标记策略，并进行了更长时间的训练，以改善结果同时保留原有的美学特征。 !!! note SHA256 c1799d22a355ba25c9ceeb6e3c91fc61788c8e274b73508ae8a15877c5dbcf63

Web

[https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step95000.ckpt](https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step95000.ckpt)

trinart2_step60000.ckpt [6ecd8e48]

V2 checkpoint使用了Dropout、多了1万张图像以及采用了新的标签策略，训练时间更长，以提高结果并保持原有的美学风格。 !!! note SHA256 0acc70d6a515cdedbc840e2c06525709a83cb9be8cdfdf1136a6548e9ee8f0fa

Web

[https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step60000.ckpt](https://huggingface.co/naclbit/trinart_stable_diffusion_v2/resolve/main/trinart2_step60000.ckpt)

trinart_stable_diffusion_epoch3.ckpt [9d7f05fc]

这个模型已经被认为是过时的了，请参考trinart2。 trinart_stable_diffusion是一个SD模型，经过大约3.5个epoch的微调，使用了约30,000张高分辨率的漫画/动画风格的图片。 !!! note SHA256 840bbc99a603d4171963b31eec700abcb6f9545d1374b53eb7f830261494be19

Web

[https://huggingface.co/naclbit/trinart_stable_diffusion/resolve/main/trinart_stable_diffusion_epoch3.ckpt](https://huggingface.co/naclbit/trinart_stable_diffusion/resolve/main/trinart_stable_diffusion_epoch3.ckpt)

Furry

Furry_epoch4.ckpt [323f8dd8]

训练数据来自于 e621 的 300k 张图片。 Tag counts: [https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE](https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE) orhttps://pixeldrain.com/u/FQwRjyyk`

!!! note MD5 f8ef45a295ef4966682f6e8fc2c6830d SHA256 4160c57f98f1727f5a52cba8c844656fc9061311f9c37daf45e8e0ebe913c987

Torrent

magnet:?xt=urn:btih:a9635389ae4c5583b0cc76ec8f6dce35438b3016&dn=furry_epoch4.ckpt&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2F9.rarbg.com%3A2810%2Fannounce&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A6969%2Fannounce&tr=udp%3A%2F%2Fopentracker.i2p.rocks%3A6969%2Fannounce&tr=https%3A%2F%2Fopentracker.i2p.rocks%3A443%2Fannounce&tr=http%3A%2F%2Ftracker.openbittorrent.com%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.torrent.eu.org%3A451%2Fannounce&tr=udp%3A%2F%2Fopen.stealth.si%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.tiny-vps.com%3A6969%2Fannounce&tr=udp%3A%2F%2Fzecircle.xyz%3A6969%2Fannounce&tr=udp%3A%2F%2Fyahor.ftp.sh%3A6969%2Fannounce&tr=udp%3A%2F%2Fvibe.sleepyinternetfun.xyz%3A1738%2Fannounce&tr=udp%3A%2F%2Fv2.iperson.xyz%3A6969%2Fannounce&tr=udp%3A%2F%2Fuploads.gamecoast.net%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker2.dler.org%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker1.bt.moack.co.kr%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.theoks.net%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.tcp.exchange%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.swateam.org.uk%3A2710%2Fannounce&tr=udp%3A%2F%2Ftracker.publictracker.xyz%3A6969%2Fannounce&tr=http%3A%2F%2Ftracker.bt4g.com%3A2095%2Fannounce

Web

[https://pixeldrain.com/u/dtYiYN7g](https://pixeldrain.com/u/dtYiYN7g)

Web

[https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch4.ckpt](https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch4.ckpt)

Furry_epoch1.ckpt [0c891127]

训练数据来自于 e621 的 300k 张图片。 Tag counts: [https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE](https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE) orhttps://pixeldrain.com/u/FQwRjyyk`

Torrent

magnet:?xt=urn:btih:d62bc9a088b206565005cab915a58fd26da1802e&dn=furry_epoch1.ckpt&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2F9.rarbg.com%3A2810%2Fannounce&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A6969%2Fannounce&tr=udp%3A%2F%2Fopentracker.i2p.rocks%3A6969%2Fannounce&tr=https%3A%2F%2Fopentracker.i2p.rocks%3A443%2Fannounce&tr=http%3A%2F%2Ftracker.openbittorrent.com%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.torrent.eu.org%3A451%2Fannounce&tr=udp%3A%2F%2Fopen.stealth.si%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.tiny-vps.com%3A6969%2Fannounce&tr=udp%3A%2F%2Fzecircle.xyz%3A6969%2Fannounce&tr=udp%3A%2F%2Fyahor.ftp.sh%3A6969%2Fannounce&tr=udp%3A%2F%2Fvibe.sleepyinternetfun.xyz%3A1738%2Fannounce&tr=udp%3A%2F%2Fv2.iperson.xyz%3A6969%2Fannounce&tr=udp%3A%2F%2Fuploads.gamecoast.net%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker2.dler.org%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker1.bt.moack.co.kr%3A80%2Fannounce&tr=udp%3A%2F%2Ftracker.theoks.net%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.tcp.exchange%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.swateam.org.uk%3A2710%2Fannounce&tr=udp%3A%2F%2Ftracker.publictracker.xyz%3A6969%2Fannounce&tr=http%3A%2F%2Ftracker.bt4g.com%3A2095%2Fannounce

Web

[https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch1.ckpt](https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch1.ckpt)

Furry_epoch0.ckpt [8c19ee5a]

训练数据来自于 e621 的 300k 张图片。 Tag counts: [https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE](https://mega.nz/file/co0UlQ5Z#vERcoYTWGJguTsXmysbLq1NL_xBS8txQhVvPI5E3QKE) orhttps://pixeldrain.com/u/FQwRjyyk`

Web

[https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch0_ckpt](https://iwiftp.yerf.org/Furry/Software/Stable%20Diffusion%20Furry%20Finetune%20Models/Finetune%20models/furry_epoch0_ckpt)

Cal Arts Style

信息和样本 [https://publicprompts.art/cal-arts-dreambooth-model/](https://publicprompts.art/cal-arts-dreambooth-model/) 触发短语： CALARTS

Web

[https://huggingface.co/PublicPrompts/CALARTS/resolve/main/CALARTS.ckpt](https://huggingface.co/PublicPrompts/CALARTS/resolve/main/CALARTS.ckpt)

Pixel Landscapes V1

信息和示例请访问 [https://publicprompts.art/pixel-landscapes-v1-dreambooth-model/](https://publicprompts.art/pixel-landscapes-v1-dreambooth-model/) 触发短语：16-bit-landscape pixel art style

Web

[https://huggingface.co/PublicPrompts/16-bit-landscape_PublicPrompts/resolve/main/16-bit-landscape_PublicPrompts.ckpt](https://huggingface.co/PublicPrompts/16-bit-landscape_PublicPrompts/resolve/main/16-bit-landscape_PublicPrompts.ckpt)

Web

[https://drive.google.com/file/d/113V4jfwlgCsAKiMLPgpj0hfdZk1sV5cL/view](https://drive.google.com/file/d/113V4jfwlgCsAKiMLPgpj0hfdZk1sV5cL/view)

MicroWorlds

信息和示例请访问 https://publicprompts.art/microworlds-dreambooth-model/ 触发短语：: microworld render style

Web

[https://mega.nz/file/7YJRFbxS#FmTrxB4ayFxUaBz-wiVxHov3YLAUFhObR2JyVwq_p_o](https://mega.nz/file/7YJRFbxS#FmTrxB4ayFxUaBz-wiVxHov3YLAUFhObR2JyVwq_p_o)

Web

[https://drive.google.com/file/d/1QTUATU3WbYIWX-6ECRypCi1z9TOsXIhG/view?usp=sharing](https://drive.google.com/file/d/1QTUATU3WbYIWX-6ECRypCi1z9TOsXIhG/view?usp=sharing)

App Icons Generator V1

信息和示例请访问 https://publicprompts.art/app-icons-generator-v1-dreambooth-model/ 触发短语: SKSKS app icon

Web

[https://drive.google.com/file/d/1TNZTQfk0CNZg7nwy033olwOpQAkcglAN/view?usp=sharing](https://drive.google.com/file/d/1TNZTQfk0CNZg7nwy033olwOpQAkcglAN/view?usp=sharing)

Pixel Art V1

信息和示例请访问 https://publicprompts.art/pixel-art-v1-dreambooth-model/ 触发短语: SKSKS art style

Web

[https://drive.google.com/file/d/1HwiqDNm3FyxMNEZLqh7FXsMJv9wmy9bc/view?usp=sharing](https://drive.google.com/file/d/1HwiqDNm3FyxMNEZLqh7FXsMJv9wmy9bc/view?usp=sharing)

VTT RPG

一个与D&D种族和怪物有关的模型集合。信息和下载链接：[https://huggingface.co/VTTRPGResources](https://huggingface.co/VTTRPGResources)

Comic Diffusion

代码库 [https://huggingface.co/ogkalu/Comic-Diffusion](https://huggingface.co/ogkalu/Comic-Diffusion) 包括了 6 种漫画风格。

The tokens for V2 are: charliebo artstyle holliemengert artstyle marioalberti artstyle pepelarraz artstyle andreasrocha artstyle jamesdaly artstyle

Web

[https://huggingface.co/ogkalu/Comic-Diffusion/resolve/main/comic-diffusion-V2.ckpt](https://huggingface.co/ogkalu/Comic-Diffusion/resolve/main/comic-diffusion-V2.ckpt)

SD_PixelArt_SpriteSheet_Generator

这个模型可以生成四个不同角度的像素艺术精灵表。仓库链接为 [https://huggingface.co/Onodofthenorth/SD_PixelArt_SpriteSheet_Generator](https://huggingface.co/Onodofthenorth/SD_PixelArt_SpriteSheet_Generator)。

“PixelartFSS” 用于生成前视图， “PixelartRSS” 用于生成右视图， “PixelartBSS” 用于生成后视图， “PixelartLSS” 用于生成左视图。

Web

[https://huggingface.co/Onodofthenorth/SD_PixelArt_SpriteSheet_Generator/resolve/main/PixelartSpritesheet_V.1.ckpt](https://huggingface.co/Onodofthenorth/SD_PixelArt_SpriteSheet_Generator/resolve/main/PixelartSpritesheet_V.1.ckpt)

midjourney-v4-diffusion

这是在Midjourney v4图像上进行微调的稳定扩散模型。使用提示：mdjrny-v4 style 存储库链接：[https://huggingface.co/prompthero/midjourney-v4-diffusion](https://huggingface.co/prompthero/midjourney-v4-diffusion)

Web

[https://huggingface.co/prompthero/midjourney-v4-diffusion/resolve/main/mdjrny-v4.ckpt](https://huggingface.co/prompthero/midjourney-v4-diffusion/resolve/main/mdjrny-v4.ckpt)

BloodborneDiffusion

这是一个基于Bloodborne系列风格的Dreamboothed Stable Diffusion模型。总数据集由100张图片组成，并使用runawayml 1.5和新的VAE进行了训练，训练步数为12k步（poly LR1e-6）。TokenBloodborne Style将引入新的概念。建议采用k_Euler_a或DPM++ 2M Karras上的20步、CFGS 7进行抽样。Repo [https://huggingface.co/Guizmus/BloodborneDiffusion](https://huggingface.co/Guizmus/BloodborneDiffusion)

Web

[https://huggingface.co/Guizmus/BloodborneDiffusion/resolve/main/BloodborneStyle-v1.ckpt](https://huggingface.co/Guizmus/BloodborneDiffusion/resolve/main/BloodborneStyle-v1.ckpt)

samdoesarts_style [85b77ff9]

这是一个基于 Sam Yang（samdoesarts）的作品训练的 Dreambooth 模型。使用 automatic1111 的 Dreambooth 扩展进行训练， 13000 步，128 张训练图片（256 翻转），1500 张分类图片，1e-6 的学习率。该模型的标记为 samdoesarts style。

Torrent

magnet:?xt=urn:btih:7999b19ee45d505354e50c25041f405ca8932a38&dn=samdoesarts_style.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2ffe.dealclub.de%3a6969%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=udp%3a%2f%2fbt1.archive.org%3a6969%2fannounce&tr=udp%3a%2f%2f6ahddutb1ucc3cp.ru%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce

Web

[https://anonfiles.com/JbocEcG4yd/samdoesarts_style_ckpt](https://anonfiles.com/JbocEcG4yd/samdoesarts_style_ckpt)

Web

[https://pixeldrain.com/u/4s3ntWvR](https://pixeldrain.com/u/4s3ntWvR)

JWST Deep Space Diffusion

这是一个基于v1.5的Fine-tuned Stable Diffusion模型，使用詹姆斯·韦伯太空望远镜拍摄的图像和Judy Schmidt的图像进行训练。在提示中使用标记JWST来使用该样式（例如，jwst, green spiral galaxy）。 Repo [https://huggingface.co/dallinmackay/JWST-Deep-Space-diffusion](https://huggingface.co/dallinmackay/JWST-Deep-Space-diffusion)

Web

[https://huggingface.co/dallinmackay/JWST-Deep-Space-diffusion/resolve/main/JWST-Deep-Space.ckpt](https://huggingface.co/dallinmackay/JWST-Deep-Space-diffusion/resolve/main/JWST-Deep-Space.ckpt)

copeseethemaldchinai_10000.ckpt (samdoesart) [32186669]

这是另一个基于Samdoesart作品的Dreambooth模型。

100 images No class images. 1e-6 learning rate. Sample guidance scale set to 9.5 based on "chinai (anything)" Use copeseethemald style in prompt

100张图片无分类图片。 1e-6的学习率。样本指导尺度设置为9.5 基于“chinai(任何东西)” 在提示中使用copeseethemald style

Torrent

magnet:?xt=urn:btih:4f1365b3dd30541d2c1becfd5d54160928916ba4&dn=copeseethemaldchinai_10000.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2ffe.dealclub.de%3a6969%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=udp%3a%2f%2fbt1.archive.org%3a6969%2fannounce&tr=udp%3a%2f%2f6ahddutb1ucc3cp.ru%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce

Web

[https://mega.nz/file/xT9jVToK#Sj1S76kl-PC-zCRwJ2FWen6DS0NHY0IXFFAkXhm03eo](https://mega.nz/file/xT9jVToK#Sj1S76kl-PC-zCRwJ2FWen6DS0NHY0IXFFAkXhm03eo)

CopeSeetheMald-berry200_20400.ckpt (samdoesart) [fa49a214]

Another samdoesart dreambooth model.

204 images @ 20.4k steps 1e-6 learning rate based on "berry mix" Use copeseethemald style in prompt

Torrent

magnet:?xt=urn:btih:e4f1f4b9d7d6cf8570914fe22318b44c18d6d602&dn=CopeSeetheMald-berry200_20400.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2ffe.dealclub.de%3a6969%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=udp%3a%2f%2fbt1.archive.org%3a6969%2fannounce&tr=udp%3a%2f%2f6ahddutb1ucc3cp.ru%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce

Web

[https://mega.nz/folder/1a1xkQQK#4atlB1cJqI35InXxlxyA7A](https://mega.nz/folder/1a1xkQQK#4atlB1cJqI35InXxlxyA7A)

CopeSeetheMald-200_20400.ckpt (samdoesart) [95f071f9]

Another samdoesart dreambooth model.

204 images @ 20.4k steps 1e-6 learning rate based on "blossom mix" Use copeseethemald style in prompt

Torrent

magnet:?xt=urn:btih:9b92aeea5d136ddae25c0d6165405e6e4d144ff1&dn=CopeSeetheMald-200_20400.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2ffe.dealclub.de%3a6969%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=udp%3a%2f%2fbt1.archive.org%3a6969%2fannounce&tr=udp%3a%2f%2f6ahddutb1ucc3cp.ru%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce

Web

[https://mega.nz/folder/ZG0UnRBJ#jykESWBUCr7hjOoNVTXwLw](https://mega.nz/folder/ZG0UnRBJ#jykESWBUCr7hjOoNVTXwLw)

Black Souls bs_1500.ckpt [37ec0dc9]

A dreambooth trained on game cg from a game called black souls call forbs in the prompt have vae off, ensure clip skip is set to 1, and set cfg to a low number (like 6) Preview https://cdn.discordapp.com/attachments/1038283286046322819/1040125269761081444/grid-0000.png

Torrent

magnet:?xt=urn:btih:63e9c043838f7bca078fb9607873957d11a88e24&dn=bs_1500.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.pomf.se%3a80%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.altrosky.nl%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2ffe.dealclub.de%3a6969%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=udp%3a%2f%2fbt1.archive.org%3a6969%2fannounce&tr=udp%3a%2f%2f6ahddutb1ucc3cp.ru%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce

Web

[https://mega.nz/file/KkIgRY7a#VrKp_JwspTF4e05ApSy4hkJgGUcMKaXSbyB_COZ-o6k](https://mega.nz/file/KkIgRY7a#VrKp_JwspTF4e05ApSy4hkJgGUcMKaXSbyB_COZ-o6k)

jaggy92500.ckpt [93423a00]

游戏截图.

这里是调用的术语：jaggy、ps_game_screenshot、pc_game_screenshot、gamecube_game_screenshot、blood_and_gore、intense_violence、zotov_game_screenshot、eroge_game_screenshot、life_sim_game_screenshot和/或alter_echo。这是按影响力降序排列的。还有一些其他的术语不足以命名，但有时会产生强烈的影响。我不会提到它们。您应该在否定提示中加入(cartoon)和可能的(menu)。预览图像：[https://imgur.com/a/J4i1DiN](https://imgur.com/a/J4i1DiN)

Torrent

magnet:?xt=urn:btih:71847b56518e6d7d92dfe24e95f2ee474807b197&dn=jaggy92500.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.lilithraws.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.cyber-hub.net%3a443%2fannounce

Web

[https://mega.nz/file/0SER2YpC#_MRc6p_sG9cSWqihpt33jpOWyMR8bCZrUaVkh4z5kGE](https://mega.nz/file/0SER2YpC#_MRc6p_sG9cSWqihpt33jpOWyMR8bCZrUaVkh4z5kGE)

ykgl.ckpt (y2k cgi girls)

It does cgi girls from the y2k era. Trained for 40k steps. You call on them with (ykgl cgi_girl), or (ykgl cgi_girls), or just (ykgl girl), and then maybe with , cgi_artstyle. Preview images https://imgur.com/a/r2aIcxo

Web

[https://mega.nz/file/hT0mgTqR#d8g133APl30UtDwsNmzV73_ZESi_kTa5pmQgJoxomn0](https://mega.nz/file/hT0mgTqR#d8g133APl30UtDwsNmzV73_ZESi_kTa5pmQgJoxomn0)

CSRmodel (cutesexyrobutts) [b77538cc]

Dreambooth model based on cutesexyrobutts art style. Preview https://i.imgur.com/VPNUae8.png

Torrent

magnet:?xt=urn:btih:46c46d2630043f9e47b0107b9680b5bd270263d7&dn=Systemy-CSRmodel.rar&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2ftracker1.bt.moack.co.kr%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.dler.org%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.lilithraws.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.cyber-hub.net%3a443%2fannounce

Web

[https://gofile.io/d/D1L69E](https://gofile.io/d/D1L69E)

Pepestyle

A Dreambooth model created with the sole purpose of generating the rarest and dankest pepes. StableDiffusion 1.5 was used as a base for this model. 22 instance images, 400 class images, 2.2k steps at a 1.3e-6 learning rate. Use the phrase pepestyle person Repo https://huggingface.co/SpiteAnon/Pepestyle

Web

[https://huggingface.co/SpiteAnon/Pepestyle/resolve/main/pepestylev2_2200.ckpt](https://huggingface.co/SpiteAnon/Pepestyle/resolve/main/pepestylev2_2200.ckpt)

dbmai [e02601f3]

A dreambooth model from China. Preview images https://imgur.com/a/fTDgBST Info https://tieba.baidu.com/p/8136937175

Torrent

magnet:?xt=urn:btih:89ede059b4f90887b1f1ce10ba7a37e1d8a0c4b5&dn=dbmai.ckpt&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce&tr=udp%3a%2f%2fopen.tracker.cl%3a1337%2fannounce&tr=udp%3a%2f%2ftracker.torrent.eu.org%3a451%2fannounce&tr=udp%3a%2f%2fopentracker.i2p.rocks%3a6969%2fannounce&tr=udp%3a%2f%2fopen.stealth.si%3a80%2fannounce&tr=udp%3a%2f%2f9.rarbg.com%3a2810%2fannounce&tr=https%3a%2f%2fopentracker.i2p.rocks%3a443%2fannounce&tr=http%3a%2f%2ftracker.openbittorrent.com%3a80%2fannounce&tr=udp%3a%2f%2ftracker.tiny-vps.com%3a6969%2fannounce&tr=udp%3a%2f%2ftracker.moeking.me%3a6969%2fannounce&tr=udp%3a%2f%2fp4p.arenabg.com%3a1337%2fannounce&tr=udp%3a%2f%2fopen.demonii.com%3a1337%2fannounce&tr=udp%3a%2f%2fmovies.zsw.ca%3a6969%2fannounce&tr=udp%3a%2f%2fipv4.tracker.harry.lu%3a80%2fannounce&tr=udp%3a%2f%2fexplodie.org%3a6969%2fannounce&tr=udp%3a%2f%2fexodus.desync.com%3a6969%2fannounce&tr=https%3a%2f%2ftracker.nanoha.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.lilithraws.org%3a443%2fannounce&tr=https%3a%2f%2ftracker.cyber-hub.net%3a443%2fannounce&tr=https%3a%2f%2ftr.burnabyhighstar.com%3a443%2fannounce

Web

[https://drive.google.com/drive/folders/1MUtM5MTXM35fid5TE3tUSKXAqPP7oysz](https://drive.google.com/drive/folders/1MUtM5MTXM35fid5TE3tUSKXAqPP7oysz)

gigachad-diffusion

StableDiffusion v1.5 used. 20 instance images, 400 class images, 2.2k steps at a 1e-6 learning rate. Token gigachad Info & preview images: https://huggingface.co/SpiteAnon/gigachad-diffusion

Web

[https://huggingface.co/SpiteAnon/gigachad-diffusion/blob/main/gigachad_2000.ckpt](https://huggingface.co/SpiteAnon/gigachad-diffusion/blob/main/gigachad_2000.ckpt)

Complex-Lineart

Trained on around 100 images at 768x768 resolution. Use prompt: ComplexLA style Use resolution near 768x768, lower resolution works but quality will not be as good. Repo & preview images: https://huggingface.co/Conflictx/Complex-Lineart

Web

[https://huggingface.co/Conflictx/Complex-Lineart/resolve/main/ComplexLA%20Style.ckpt](https://huggingface.co/Conflictx/Complex-Lineart/resolve/main/ComplexLA%20Style.ckpt)

		自动登录	找回密码
密码			立即注册

[教程] Stable Diffusion 大模型资源下载集合